Reckless Complicity Sanford H

Journal of Criminal Law and Criminology Volume 87 Article 1 Issue 2 Winter

Winter 1997 Reckless Complicity Sanford H. Kadish

Follow this and additional works at: https://scholarlycommons.law.northwestern.edu/jclc Part of the Criminal Law Commons, Criminology Commons, and the Criminology and Criminal Justice Commons

Recommended Citation Sanford H. Kadish, Reckless Complicity, 87 J. Crim. L. & Criminology 369 (1996-1997)

This Criminal Law is brought to you for free and open access by Northwestern University School of Law Scholarly Commons. It has been accepted for inclusion in Journal of Criminal Law and Criminology by an authorized editor of Northwestern University School of Law Scholarly Commons. 0091-4169/96/87020369 THEJOURNAL OF CRIMINA LAW & CRIMINOLOGY Vol. 87, No. 2 Copyright @ 1997 by Northwestern University, School of Law Printed in U.S.A.

CRIMINAL LAW

RECKLESS COMPLICITY

SANFORD H. KADISH*

The law of complicity (or accomplices or aiding and abetting, as it is sometimes called) imposes liability on one person ("S" for secondary party) for a crime committed by another ("P" for principal party) where S intentionally helps or encourages P to do the action constituting that crime. The law is inescapably complex since it involves two dimensions of interacting mens rea and actus reus problems, those associated with S and those associated with P. I attempted to deal with complicity law generally over a decade ago.' In this paper I revisit one group of issues, namely those arising from the mens rea requirement that S must intend his actions to encourage or help P to commit a particular crime. My earlier study was concerned mainly with why the law of complicity should have taken the form it did. Here I want to take a more critical view of the requirement of intention and ask how far it can withstand critical scrutiny. The puzzle in the requirement of intention is that it stands in contrast with situations where the accountability of an actor for the harmful consequences of his action turns on whether he caused them. In these latter cases, recklessness as to the occurrence of the consequences (and sometimes negligence) is enough to make the actor criminally liable-the crime of manslaughter is the commonest example. Why should it be different where the consequences of a person's action take the form of the criminal actions of another? Can reckless

* Morrison Professor of Law, Emeritus, University of California, Berkeley. Aversion of this paper was presented at the Institute of Criminal Law at Tel Aviv University and at the Max-Planck Institute of Foreign and International Criminal Law at the University of Frei- burg. I am grateful to the participants for their many helpful comments as well as to those friends and colleagues who generously read and responded to an earlier draft: Andrew Ashworth, Meir Dan-Cohen, Joshua Dressler, Claire Finkelstein, Christopher Kutz, Mort Kadish, Herbert Morris, Eric Rakowski, Steven Schulhofer and Philip Selznick. 1 Sanford H. Kadish, Complicity, Cause and Blame: A Study in the Interpretationof Doctrine, 73 CAUF. L. REv. 323 (1985).

369 370 SANFORD H. KADISH [Vol. 87

aiding or encouraging the crime of another ever be an acceptable basis for holding S criminally liable? These are the questions I will pur- sue in this paper. The thread of my argument is as follows. Part I argues that three main considerations lie behind the law's general requirement that S's help or encouragement of P's crime be intentional-culpability, the policy of not subjecting lawful practices to excessive risk, and an ethic of individualism and self-determinism. The remainder of the paper considers whether those considerations are incompatible with recklessness as an alternative basis of complicity liability. Parts II and III deal with the culpability consideration. Part II describes the one situation where English and American law systematically depart from the requirement of intention and allow negligence or even less to suffice. This occurs in what the English misleadingly call "common purpose" or 'Joint enterprise" cases, where S is already an accomplice of P in some other crime (misleadingly, because the common purpose is to commit some other crime, not the one P commits). I argue that the major defects of this doctrine are that it permits punishment for mere negligence, or even less, and that it provides for the conviction of S for a crime for which S lacks the required culpability or its equivalent. In Part III, I argue that the doctrine of reckless complicity can be made compatible with the requirement of culpability by restricting it to cases where P's crime is one of recklessness or can be committed recklessly. I then consider in Part IV whether the policy of not chilling lawful activities precludes the creation of a general reckless complicity principle untied to the situation where S is already an accomplice of P in some other crime. I argue that it does not, and therefore conclude that neither the principle of culpability nor policy considerations require rejection of a general doctrine of reckless complicity. The final section, Part V, considers the ethic of individualism and self-determinism, and concludes that it does indeed stand against reckless complicity as a general ground of liability, and is the real force behind the law's requirement of intention; and further that this ethic, while deeply ingrained, is normatively problematic. A word about what I am up to in all this. Although I later suggest how the law might be modified to embrace a doctrine of reckless complicity, it is no part of my purpose to advocate that this be done. To announce at the outset what I later will want to repeat, I would find unwise any radical extension of the reach of the criminal law in view of the realities of the administration of criminal justice, at least in America. My purpose here is only to try to get at the underlying reasons why American and English law, with the exception of the common purpose cases, have insisted on the requirement of intention. 1997] RECKLESS COMPLICITY

I. THE REQUIREMENT OF INTENTION

I begin with the theory behind the prevailing requirement of in- tendon (sometimes broadly interpreted to include knowledge, as we shall see) and suggest three considerations that seem to lie behind it. The first is culpability, the second the policy of safeguarding lawful conduct from risk, and the third an ethic of individualism and self- determinism. The culpability consideration arises from the need that S satisfy the culpability requirement of P's crime in order that he may be found guilty of it. For example: S reveals the combination of his employer's safe to P, whom he knows to be a career burglar. Say he does so because he is bedridden, and it is therefore convenient for him to have P fetch some records he needed from his employer's safe. But say also that he is fully aware of a substantial risk that it will enable P at any time in the future to open the safe and steal its contents; for him it is worth that risk in order to get the papers. On a later occasion P breaks into the premises, opens the safe and steals cash stored there. P, of course, has the mens rea of burglary-he entered with intent to commit the felony of theft, that is, to appropriate another's property with the intention permanently to deprive the owner of it. Did S have that mens rea? Yes, if S gave P the combination with the intention of helping P commit the burglary. Not precisely, it must be said, since burglary requires that the defendant do the breaking and entering with the required intent, and here S intended P to do so, not to do so himself. But if S helps P with the intention that P should do so S intends to participate in P's burglary and therefore may fairly be regarded as having a culpability equivalent to that of P. In our hypothetical, however, since S only acted recklessly with regard to the risk that P might burglarize the premises he can not be said to have intended to participate in the burglary and therefore did not act with a culpability equivalent to that of P. In sum, the requirement that S intentionally help P to commit the crime assures that S acts with a culpability equivalent to that which P's crime requires. The policy concern is that to burden peoples' actions with doubts and worries about what someone else might culpably do as a consequence of their own lawful actions would tend to create an undesir- able insecurity in the conduct of ordinary affairs. It is true that the law commonly imposes on people the burden of avoiding unwarranted risks that their conduct will cause harm, at least when they are aware of the danger and the harm eventuates-manslaughter, for example. But it is apparently believed that a burden of avoiding unwarranted risks that their conduct will help or encourage another to commit crime SANFORD H. KADISH [Vol. 87

is distinguishable in nature and degree. 2 Finally, there is the ethic of individualism and self-determinism reflected in traditional criminal law doctrine that except in limited specified circumstances neither what happens to another nor what another person does is one's responsibility. I will discuss the relevance of this ethic to the intention requirement in a later section of this paper.

II. UNINTENTIONAL COMPLiCITY-AN EXCEPTION Do these three considerations always require that S act intentionally? Is there ever a case for holding a non-intentional S liable for P's crime? I begin with the one situation where Anglo-American law most clearly dispenses with the requirement of intention and holds S liable for P's crime even though S neither knew nor intended that his conduct would help or encourage it. This the law does in cases of so- called "common purpose," where S is already on the hook for intentionally assisting P to commit some other crime. The law's inconsistency in not requiring intention here while at the same time insisting upon it in other situations is illustrated in a decision of the California Supreme Court, People v. Beeman.3 In Beeman, the defendant was convicted as an aider and abettor of two men who burglarized the home of his sister-in-law. There was evidence that he informed them of the contents and layout of her home knowing they intended to commit the burglary. His defense was that he provided the help innocently, that is, without intending thereby to facilitate the burglary. The trial court instructed the jury that it was enough to convict him that he provided the help "with knowledge of the unlawful purpose of the perpetrator of the crime." But the supreme court found this to be error and reversed his conviction, holding that "the weight of authority and sound law require proof that an aider and abettor act with knowledge of the criminal purpose of the perpetrator and with an intent or purpose either of committing, 4 or of encouraging or facilitating commission of, the offense." This requirement of an actual purpose constitutes a strict reading of the intent requirement; some jurisdictions accept the looser view of intention that makes knowledge sufficient.5 Yet even a jurisdiction

2 These concerns parallel those commonly expressed for limiting criminal liability for omissions. See infra text accompanying note 57. 3 674 P.2d 1318 (Cal. 1984). 4 Id. at 1325. 5 This is apparently a minority position in the United States, applied primarily in cases of serious crimes. See cases cited in Paul H. Robinson, Imputed CriminalLiability, 93 YALE L J. 609, 637 n.100 (1984). The principal argument for the majority view is that making 1997] RECKLESS COMPLICITY 373

like California that takes the stricter view of intention abandons any version of that requirement where S is already in league with P in some other criminal venture. Thus, the court in Beeman, after reciting its strict view of intention, reasserted the principle that where S intentionally aids and abets P to commit some crime, he is liable not only for that crime but for any that are the "natural and reasonable consequences of the acts he knowingly and intentionally aids and encourages."6 As restated in Croy,7 a later decision of the California Supreme Court, an aider and abettor: need not have intended to encourage or facilitate the particular offense ultimately committed by the perpetrator. His knowledge that an act which is criminal was intended, and his action taken with the intent that the act be encouraged or facilitated, are sufficient to impose liability on him for any reasonably foreseeable offense committed as a consequence by the perpetrator. It is the intent to encourage and bring about conduct that is criminal, not the specific intent that is an element of the target offense, which Beeman holds must be found by the jury. These formulas are a far cry from intention, strict or loose. Their practical application is exhibited in a case like People v. Luparellos where the defendant instructed his men to obtain certain information from deceased "at any cost." When his men failed to get the information from the deceased one of them waited for him outside his home and shot him dead. The court affirmed defendant's conviction of first degree murder (requiring premeditation and deliberation), upholding an instruction that allowed the jury to find him guilty if the murder was the natural consequence of the assault he encouraged. The court said: [T]o be a principal to a crime... the aider and abettor must intend to commit the offense or to encourage or facilitate its commission. Liabil- ity is extended to reach the actual, rather than the planned or 'intended' crime, committed on the policy [that] aiders and abettors should be responsible for the criminal harms they have naturally, probably and foreseeably put in motion.9 While I have used California decisions as illustrations, the same approach is employed in many American jurisdictions,10 in some explic-

knowing but disinterested aid sufficient for liability unduly threatens the conduct of lawful business. 6 Beeman, 674 P.2d at 560. The formula varies. In People v. Durham, 449 P.2d 198, 204 (Cal. 1969), the court spoke of "natural and probable" consequences; and in People v. Croy, 710 P.2d 392, 398 n.5 (Cal. 1985) the court spoke of "any reasonably foreseeable offense committed" by P. 7 People v. Croy, 710 P.2d at 398 n.5 (Cal. 1985). 8 231 Cal. Rptr. 832 (1987). 9 Id. at 849. 10 A classic article of Professor Sayre in 1930 describes this approach as the prevailing one. Francis Sayre, CriminalResponsibilityfor the Acts of Another, 43 HARv. L. REv. 689, 702-06 SANFORD H. KADISH [Vol. 87

itly by statute." English law appears to differ from California law in that S may be an accomplice in P's crime so long as he knows his action will help P commit a crime; intention in the sense of purpose is not always needed.12 But once S has become an accomplice of P in one crime, even knowledge of P's intention to commit a second crime is not needed-it is enough that P's second crime "was foreseen as a possible incident of the common unlawful enterprise." I s So in a recent case, Regina v. Hyde,14 where three defendants as- saulted the deceased, but the prosecution could not prove who deliv- ered the fatal blow or that grievous injury was contemplated by any when the attack was launched, the court concluded that all could be convicted of murder. At least one of them struck with intent to inflict grievous bodily harm, a sufficient mens rea for murder. Even if the other two neither struck the blow nor planned that one of them should do so, the jury could convict them all of murder upon a finding that they foresaw the danger that one of them would strike a fatal blow. So the court concluded: If B realizes (without agreeing to such conduct being used) that A may

(1930). See, more recently, WAYNE LAFAVE & AuSTIN W. Sco-r, JR., CRIMINAL LAW 590 (2d ed. 1986): "The established rule, as it is usually stated by courts and commentators, is that accomplice liability extends to acts of the principal in the first degree which were a 'natural and probable consequence' of the criminal scheme the accomplice encouraged or aided." 11 See e.g., ME. REv. STAT. ANN. tit. 17-A, § 57 (West 1995) (emphasis added): 3. A person is an accomplice of such other person in the commission of the crime if: A. With the intent of promoting or facilitating the commission of the crime, he solicits .... or aids or agrees to aid or attempts to aid such other person in planning or committing the crime. A person is an accomplice under this subsection to any crime the commission of which was a reasonablyforeseeable consequence of his conduct ... Some of these statutes require the foreseeable crime to be committed in furtherance of the intended crime. See, e.g., MINN. STAT. § 609.05 (1996) (emphasis added): Subd. 1. A person is criminally liable for a crime committed by another if the person intentionally aids, advises, hires, counsels, or conspires with or otherwise procures the other to commit the crime. Subd. 2. A person liable under subdivision 1 is also liable for any other crime committed in pursuance of the intended crime if reasonably foreseeable by the person as a probable consequence of committing or attempting to commit the crime intended. See also IowA CODE ANN. § 703.2 (1995); KAN. STAT. ANN. § 21-3205 (1995); Wis. STAT. ANN. § 939.05 (West 1995). 12 See National Coal Board v. Gamble, 1 Q.B. 11 (1959). But some qualification of that position may be required in light of Gillick v. West Noifolk and Wisbech AHA, 1986 A.C. 112. See The Law Commission, Assisting and Encouraging Crime, Consultation Paper No. 131 (1993) at 3842; K.J.M. SMITH, A MODERN TREATISE ON THE LAW OF CRIMINAL COMPLICITY 141-150 (1991). 13 Chan Wing-Siu v. The Queen, A.C. 168, 175G (P.C. 1984) (Hong Kong); Hui Chi- Ming, 94 Crim. App. 236 (P.C. 1992) (Hong Kong). The history of the development of this concept in England is told in K.J.M. SMITH, supra note 12, at 209-34. Australian law is in accord. Jones v. The Queen, 143 C.L.R. 108 (1980); BRENT FISSE, HowARD's CRIMINAL LAw 338 (5th ed. 1990). 14 1 Q.B. 134 (1991). 1997] RECKLESS COMPLIMTY S375

kill or intentionally inflict' serious injury, but nevertheless continues to participate with A in the venture, that will amount to a sufficient mental element for B to be guilty of murder if A, with the requisite intent, kills 15 in the course of the venture. The court made explicit that, contrary to some indications in earlier cases, a tacit agreement or authorization need not be found; foresight 6 of the risk alone would suffice.' This common purpose doctrine has two defects. The first has to do with the degree of risk required to make S liable. The American formulation makes it suffice that the risk was reasonably foreseeable, in other words, that S was negligent. The English cases usually find it enough that S is aware of only "some" risk that P will commit the crime, suggesting that both American and English law share the defect of imposing criminal punishment for mere negligence of the kind sufficient for civil liability, or even less. The second defect, which the doctrine shares with a cognate doctrine in the law of conspiracy,' 7 has been widely noted 8-it serves to convict S of the crime of P even where he acts without the culpability P's crime requires. In Hyde, for example, the defendant was convicted of murder even though he intended neither himself nor another to inflict grievous bodily harm. It was enough that he foresaw the possibility that one of his partners would inflict such harm. In Luparello, the defendant was convicted of a first degree premeditated killing though he neither premeditated nor in any sense intended the killing. It was enough that he was negligent as to the possibility that one of his men would kill. I suggested that the traditional requirement of intention for accomplice liability could be understood in part as serving to assure that S shares the culpability required for P's crime, or more precisely an equivalent culpability. But under the common purpose doctrine, so long as S is an accomplice in any crime of P he is thereby made an accomplice of any other crime that he should have foreseen (the American version) or did foresee (the English version) P might commit. The doctrine has more than a family resemblance to the so-

15 Id. at 139D. 16 See K.J.M. SMIH, supra note 12, at 220-21. 17 In American conspiracy law it takes the form of the so-called Pinkerton doctrine, which makes a conspirator liable for the crimes committed by another conspirator that are foreseeable and committed in furtherance of the aims of the conspiracy. Pinkerton v. United States, 328 U.S. 640 (1946). The "in furtherance" requirement constitutes a restriction that is sometimes not present in the non-conspiracy form of the doctrine. 18 For an example ofjudicia criticism, see the vigorous concurring opinion ofJustice Wiener in People v. Luparello, 231 Cal. Rptr. 832 (1987). For typical commentator criticism, see WAYNE R. LAFAVE & AUSTN W. Sco-r, JR., CRIMINAL LAw 59091 (2d ed. 1986). SANFORD H. KADISH [Vol. 87

called lesser-crime doctrine sometimes seen in American and English cases, according to which if a defendant in the course of committing crime A does an action that is proscribed by crime B, he is also guilty of crime B even if he acted without the culpability required by the definition of crime B. 19 It also shares a resemblance to the American felony-murder rule, long since abandoned in England, which is a particular application of the lesser-crime doctrine to murder: a killing committed in the course of a felony (nowadays only certain felonies) becomes murder even if, apart from the felony, it would be manslaughter or not criminal at all. 20 Like these related doctrines, then, the common purpose doctrine is essentially arbitrary in serving to convict persons of crimes for which they lack the stipulated degree of culpability; the fact that the defendant has the culpability for some crime does not itself establish his culpability for another more serious crime. Despite this criticism of the doctrine no less a person than Profes- sor Sir John C. Smith of the University of Nottingham has argued in its defense. Commenting on the Hyde case, he defended convicting all three defendants of murder as compatible with the principle of culpability on the ground that "the person who embarks on a joint enterprise knowing that his confederate may intentionally kill is taking a deliberate risk of assisting or encouraging not merely killing but murder."21 But does risking that P will intentionally kill entail an

19 See GLANVILLE WILLIAMS, CRIMINAL LAw. THE GENERAL PART 185 (2d ed. 1961); WAYNE R. LAFAVE & AUSTIN W. SCOTT, JR., CRIMINAL LAw 410 (2d ed. 1986). The origin of this doctrine is apparently the medieval canon law doctrine "versari in re illicita imputantur omnia quae sequuntur ex delicto." (One who traffics in the illicit is responsible for all wrongs that ensue). See STEPHAN KUTrNER, KANONISTISCHE SCHULDLEHRE 185 (Citta Del Vaticano, Biblioteca Apostolica Vaticana 1935); HANS-HEINRICH JESCHEcK, LEHRBUCH DES STRAFRECHTS 235 (1988). It plainly responds to a widely shared moral viewpoint. Cf Ar- thur Conan Doyle's story, The Adventure of the Prioy Schoo, involving the murder of an innocent person by a ruffian while kidnapping the Duke's younger son at the behest of the Duke's elder son. Speaking of the liability of the elder son, who meant no harm to the deceased, Holmes stated: "I must take the view, your Grace, that when a man embarks upon a crime, he is morally guilty of any other crime which may spring from it." The Duke: "Morally, Mr. Holmes. No doubt you are right. But surely not in the eyes of the law. A man cannot be condemned for a murder at which he was not present, and which he loathes and abhors as much as you do." Sm ARTHUR CONAN DOYLE, THE RETURN OF SHER- LOCK HOLMES 131 (Berkley Medallion ed. 1963). Under the common purpose doctrine, of course, his elder son would be legally liable as well. 20 LAFAvE & ScoTr, supra note 10, at 620; Lloyd Weinreb, Homicide: Legal Aspects, in ENCYCLOPEDIA OF CRIME AND JUSTICE 855, 859 (1983). 21 Case and Comment, 1991 CRIM. L. REV. 134-35: The law.., is criticised on the ground that it leaves the accessory liable to conviction for murder although he is only reckless whereas, in the case of the principal offender, intention must be proved. But there is a difference between (i) recklessness whether death be caused (which is sufficient for manslaughter but not for murder) and (ii) recklessness whether murder be committed. The accessory's recklessness must extend 1997] RECKLESS COMPLICITY 377 equivalent culpability to intentionally killing, as Smith's argument suggests? The answer from a consequentialist perspective would seem to be no. So long as the magnitude of the risk is the same the person who recklessly risks the death of a person by murder is no more culpable than one who risks his death by accident-the upshot of his action in either case is a death he did not intend. Consider two cases. In the first case S, the coach of a pistol shooting team, recklessly chooses a dangerous area for the team's target practice, reckless because it is very near a pedestrian trail and she knows it. (Though Smith's ration- ale would not seem to require an independent criminal purpose, let's assume there is one, just to bring the case within the common purpose doctrine-say, for example, S and her students are test firing guns in preparation for an illegal arms shipment.) A bullet shot by one of her students, P, who is also aware of the danger to passersby, accidentally strikes and kills one. It's clear that if P is guilty of manslaughter so is S, since she intentionally helped and encouraged P to engage in precisely the dangerous action that caused the accidental death.22 Now consider a second case. Coach S' chooses a reasonably safe area for target practice, but she acts in reckless disregard of a known substantial risk that P' will seize the opportunity to intentionally kill X, who S' knows regularly visits the area and whose life P' has threatened. P' spots X and shoots him dead. The risk taken by S and S' is equally unjustified, and if we assume that the probability of the passerby being shot by accident in the first case and by P's design in the second case is also the same, then it would seem that the culpability of coach S' should not be greater than that of S. Their equally risky and unjustified actions had the identical upshot, death of a person. But a different perspective lies behind Smith's position. On this view (call it an action perspective) consequences aren't everything; actions count too, and for their own sake, not just because of their consequences. So coach S' does a worse thing than coach S because the upshot of taking that risk is murder, and it is worse, morally worse, that is, for a person to be murdered than for a person to be killed in an accident, the reason being that murder is an action whose evil resides not simply in its upshot, but in the evil of its doing as well. For

to the principal's mens rea of murder. The person who embarks on ajoint enterprise knowing that his confederate may intentionally kill is taking a deliberate risk of assisting or encouraging not merely killing but murder .... The principle applied in these cases is not confined to murder but extends to accessory liability generally. It is not unreasonable that one who is reckless whether he assists or encourages the commission of a crime should be held liable for it when it is committed. 22 See Kadish, supra note 1, at 347; MODEL PENAL CODE § 2.06 cmts. at 321 (1985). 378 SANFORD H. KADISH [Vol. 87

the consequentialist this rests on a confusion, for what renders the action of killing evil is that it results in a death. Nonetheless, the action perspective has appeal-while the world is made a worse place when a driver kills a person by reckless driving, it is made a worse place still where a murderer intentionally does so. Therefore, in risking an intentional killing, S does worse than one who risks an accidental death, and hence is the more culpable. But even if one adopts the action perspective it means only that one who risks murder is more culpable than one who risks an accidental killing, not that she has the mens rea required for murder. We might deprecate the action of S' more strongly than that of S, but that's not to say she's as guilty as her murdering student. The student has the mens rea of murder since he intended to kill, but his coach does not, not, at any rate, simply because she risked that her student would commit murder.23 And of course the same distinction in culpability between the reckless helper and the intentional principal is evident in crimes other than murder. One who recklessly risks that another will commit burglary is not a burglar. Not possessing the intent to break and enter in order to steal, he lacks the culpability of a burglar. Nor is his culpability equivalent. His fault is risking burglary by another, not participating in a burglary.

III. A MODIFIED COMMON PURPOSE DOCTRINE- RECKLESS COMPLICITY In the end, therefore, I don't think Professor Smith's argument succeeds in saving the common purpose doctrine. But it may be sal- vaged if its two major defects are remedied. The first, it will be recal- led, is that negligently helping P, or even less, is all that is required to make S liable for P's crime. That could be dealt with by reformulating the required risk to reflect a culpability adequate to justify criminal punishment. Recklessness as defined by the Model Penal Code meets this requirement: the conscious disregard of "a substantial and unjustifiable risk" whose "disregard involves a gross deviation from the standard of conduct that a law-abiding person would observe.... "24 The second defect I described is that the doctrine derives the culpability required for the greater crime from S's guilt as an accomplice to a lesser crime committed by P. If the doctrine were modified not only to require recklessness on the part of S, but also if it were limited

23 Although she might if her recklessness was great enough to constitute what the common law called implied malice, and what is called by the Model Penal Code "[recklessness] manifesting extreme indifference to the value of human life." MODEL PENAL CODE § 210.2(1)(b) (1985). 24 MODEL PENAL CODE § 2.02(2) (c) (1985). 1997] RECKLESS COMPLiCTTY

to cases where there is a crime of reckless risk creation which S's conduct satisfies, or where P's crime is itself one of recklessness, that defect would not be present, for then S's own recklessness in helping P would supply the required culpability. My argument is that altered in these ways the common purpose doctrine would become a doctrine of reckless complicity, free of the defects that attend the present formulation of the common purpose doctrine. Let me suggest some examples of how it would work. In the Hyde case, since Hyde was reckless (let us presume) as to the risk that one of the other participants in their joint attack would go so far as to attack the victim with intent to do grievous bodily harm, he acted with a culpability equivalent to that required by the crime of manslaughter and hence could be convicted of that crime (as on this reasoning could Coach S' in my pistol team hypothetical). While Hyde is a case of recklessness as to intentional wrongdoing by P, the argument works as well where the risk recklessly taken is of some reckless action of P. The drag race cases are examples.2 5 In such cases, two drivers agree to race in the city streets, a reckless enterprise. One of them, P, kills a pedestrian in the course of the race. P is plainly guilty of manslaughter. But so is the other driver, S. It should not be an objection that S did not act with the required culpability, since by participating in the race S manifested an equivalent culpability, recklessness. It may be thought that these cases are simply applications of the traditional requirement of an intentional aiding, on the view that since reckless driving is what drag races are all about, by participating in the race each driver is intentionally helping and encouraging the other.26 But while that would explain the case of a speeding P, it would not so readily fit the case of a P who chooses in the heat of the race to take risks like driving the wrong way on a busy one-way street, or mounting a sidewalk. These are risks which S might not have intended P to take, but if he disregarded a substantial and unjustifiable risk that P might take them (say, perhaps, because he knew that P was a notoriously irresponsible driver), he has acted with the mens rea required for manslaughter.

25 E.g., State v. McFadden, 320 N.W.2d 608 (Iowa 1982); People v. Abbott, 445 N.Y.S.2d 344 (N.Y. App. Div. 1981); Commonwealth v. Root, 170 A.2d 310 (Pa. 1961). 26 Compare People v. Abbott, 445 N.Y.S.2d at 347 (upholding the manslaughter conviction of Moon for the death inflicted by his racing competitor, Abbott). In this case the court said: While Moon did not personally control Abbott's vehicle which struck and killed the three victims, it could reasonably be found that he 'intentionally' aided Abbott in the unlawful use of the vehicle by participating in a high-speed race, weaving in and out of traffic, and thus shared Abbott's culpability. The quotation marks around "intentionally" perhaps reveals the court's awareness that it was stretching a bit. 380 SANFORD H. KADISH [Vol. 87

Professor Glanville Williams disagrees that by recklessly helping P commit a crime of recklessness S manifests the recklessness required for P's crime.27 He argues this point in the context of cases in which S is not already on the hook for assisting P in some other crime, situations I defer to the next section. But if his argument holds in those cases it also holds where S is already on the hook, so I will consider his argument here. He instances a case where a father gives his car keys to his son to get gasoline, "realising that he is running a risk because the vexatious youth may seize the opportunity of taking a joy-ride," which he does.28 Williams' view is that the father can not be convicted of complicity in his son's crime of reckless driving because, first, his conduct is "in essence a reckless failure on his part to prevent the offence, which should not suffice," 29 and, second, "the recklessness about which we are now talking is recklessness in the keeping of keys, whereas the prosecution is about reckless driving."30 On Williams' first point, it seems to me that lending the car is not simply a failure to prevent the offense, it is providing the means of committing it.31 As to the second point, that different kinds of recklessness are involved, in a sense that is plainly so. The father's recklessness was in knowingly incurring the substantial and unjustifiable risk that his son, given access to the car keys, would drive recklessly; the son's recklessness was in the way he drove the car. But it is not as though the two acts of recklessness are unrelated. The son imposed a substantial and unjustifiable risk to the life and limb of others in driving as he did; so did the father in recklessly risking that his son would do so. It's true that the statute makes it a crime to drive a car recklessly, whereas the father did not drive at all, let alone drive recklessly. But, as was developed earlier, it is always true of accomplice liability that in assisting or encouraging P,S does not do the action prohibited by P's crime and hence can't be said to do it with the required mens rea; it is enough that he acts with an equivalent culpability. For example, when S gives P a weapon with which to strike a foe, S himself does not strike the blow, intentionally or otherwise. So if in the circumstances the father was reckless in giving the car keys to his son, be-

27 Glanville Williams, Complicity, Purposeand the Draft Code-2, 1990 GluM. L. Ray. 98, 99.

28 Id. at 99. 29 Id. 30 Id. 31 Id. In another hypothetical he puts the case of the father who fails to lock up his car keys, remembers in time, but is too lazy to go back, though aware of the danger posed by his son's proclivities. This case would fairly raise the issue of whether recklessly helping another to commit a crime and recklessly failing to take precautions to prevent him from doing so should be treated the same way. This issue, however, is part of the separate question of how the law should deal with liability for omissions. 1997] RECKLESS COMPLICITY

cause of the known, substantial risk that the son would engage in reckless driving, he possesses the required culpability, since he acted with a culpability equivalent to that required by the son's crime. Suppose Professor Williams's story had a more serious ending. The son, say, is notjust a vexatious youth, but also, as is well known to the father, just beginning to learn to drive, has proved singularly in- ept, and has displayed an alarming proclivity to taking wild and irresponsible risks. After buying the gasoline the son takes off on a joy ride, drives recklessly and kills someone. I would think it a mistake to deny the father's liability for manslaughter on the ground that he lacks the culpability required of that offense. He did not kill the victim and he did not drive the car recklessly. But he put the car in the hands of his son in reckless disregard of the substantial and unjustifiable risk that his son would drive recklessly and kill someone. It is not clear to me why he has not thereby displayed a culpability equivalent to that required of manslaughter. It is true that the intervention of a second actor, whose action is required for the harm to occur, reduces the probabilities that the harm would ultimately happen. One could imagine, indeed, that there might be not just one, but a chain of intervening actors. Of course with the addition of every actor the probability of the harm happening becomes less until eventually the point is reached where the risk would no longer be substantial enough to constitute recklessness. But that is not to say that the presence of a single intervening actor whose harmful action is risked necessarily precludes a finding that the risk S created was substantial and unjustifiable. Therefore, the argument from reduced probabilities does not in principle defeat the case for reckless complicity. One further point. In these father-son hypotheticals most will feel that on the facts given the father is not as blameworthy as the son. But while that would warrant a lesser sentence for the father it should not defeat his liability, any more than it defeats the liability of a secondary party in a typical case of intentional aiding that he may not be as blameworthy as the principal. My conclusion, therefore, is that a reckless complicity doctrine can not be faulted on culpability grounds so long as it is applicable only to crimes of recklessness. So the first reason I suggested for the law's general requirement of intention does not stand against the doctrine. Nor, it is evident, does the second, the policy of avoiding bur- dening otherwise lawful actions-certainly not so long as the doctrine retains the requirement that S's conduct be independently criminal, which so far I have not challenged (though I am about to). So there is no ground for concern that the doctrine would subvert the policy of 382 SANFORD H. KADISH [Vol. 87

maintaining the security of ordinary, lawful activities-the only burden is upon criminal actions. The third reason, having to do with an ethic of individualism and self-determinism, may indeed be thought to stand against the doctrine, because S, not having intentionally assisted P's crime, has arguably not exercised his will to identify with P's actions. I will return to this issue in Part V, but for the moment it may be noticed that the modified version of reckless complicity (which is the common purpose doctrine without its objectionable mens rea fea- tures) is not any less compatible with that ethic than the common purpose doctrine exhibited in such cases as Hyde and Luparello.

IV. AN ENLARGED RECKLESS COMPLICITY-THE Poucv CONSIDERATION Having made the argument for salvaging the common purpose doctrine by converting it into a reckless complicity doctrine, restricted to crimes of recklessness, the project now is to consider the case for enlarging the doctrine by dispensing with the requirement that S be already on the hook as an accomplice of P in some other crime. Does the case for reckless complicity reach all cases where S has recklessly encouraged or aided a crime of P, irrespective of whether S is already an accomplice of P in some other crime? The requirement of S's independent criminality serves two legiti- mate purposes. First, it is powerful evidence that S's creation of the risk was unjustified and therefore goes a good distance to establishing his recklessness. But this only supports the relevance of such evidence, not its necessity, for engaging in crime is obviously not the only circumstance that can make a risk unjustified. The second purpose of the requirement, however, has greater weight-it assures that liability for reckless complicity does not unduly burden the conduct of ordinary affairs, since if S's conduct is otherwise criminal there can't possi- bly be a worry that ordinary affairs would be put in jeopardy. But is the desirability for this assurance a fatal objection to the extension of liability? Certainly it is widely believed to be so. As we saw, many jurisdictions decline to make it criminal for S to render assistance to P even though S knows that P will use the help to commit a crime. These jurisdictions, like California and Model Penal Code jurisdictions, require that S act with the purpose of helping P commit the crime. The commonly given reason for this limitation is the worry that otherwise everyday lawful activities would be made perilous.32 If this is a con-

32 The Model Penal Code illustrates this concern with the following hypotheticals: "A lessor rents with knowledge that the premises will be used to establish a bordello. A vendor 1997] RECKLESS COMPLTCITY

cern with proposals to make knowing aid suffice, it is obviously a far more serious concern with proposals to make reckless aid suffice. As Professor Williams observed, it is one thing to become an accessory by helping, "knowing for a fact that crime is afoot," but quite another thing to become an accessory by helping, "knowing that a crime may be afoot."33 Even the Model Penal Code, whose original proposal (later withdrawn) favored extending liability to those who knowingly aid a criminal undertaking, did not go so far as to propose liability for reckless aid. And in England a proposal that could be construed to do so was roundly condemned by Professor Williams as absurd. He observed: The law of complicity makes me my brother's keeper, but not to the extent of requiring me to enquire whether he is engaging (or proposing to engage) in iniquity, when4 my own conduct (apart from the law of complicity) is innocent.3 Nevertheless, there are situations of reckless aid where the extension of criminal liability may be quite defensible. Consider the mother who leaves her baby in the sole care of a boyfriend she knows to have violently beaten the child on previous occasions;3 5 or the police officer who leaves his gun unattended in a neighborhood play- ground in a high crime area; 6 or my earlier examples of the person who recklessly disclosed his employer's safe combination to a known career burglar, or the pistol shooting team coach who recklessly helped her student shoot in awareness of his murderous propensity toward a person expected to be present. In such cases as these would criminal liability be an unsafe extension of the criminal law even if liability were imposed under a crime that properly reflected the actor's culpability of recklessness and that carried an appropriately mod-

sells with knowledge that the subject of the sale will be used in the commission of a crime .... A utility provides telephone or telegraph service, knowing it is used for bookmaking. An employee puts through a shipment in the course of his employment though he knows that the shipment is illegal." MODEL PENA1. CODE § 2.06 cmt. 6(c) at 316 (1985). -3 Williams, supra note 27, at 103. 34 Id. at 101. See also id. at 103: [T]he draft makes everyone an accomplice in crime, however lawful his own beha- viour, if (broadly speaking) what he did in fact assisted in an act that results in the commission of the crime, being merely aware that someone (perhaps someone whose identity was unknown to him) might do the act in question. I do not believe that this absurd rule represents the law. 35 E.g., People v. Stanciel, 606 N.E.2d 1201 (Ill. 1992); Commonwealth v. Howard, 402 A.2d 674 (Pa. Super. Ct. 1979). 36 CompareJ.C. SMITH & BRIAN Hoc, CRIMINAL LAxw 134 (7th ed. 1992): Mere recklessness, still less negligence, whether assistance be given is probably not enough. D's realization that he may have left his gun-cubbard unlocked and that his son has a disposition to commit armed robbery, is probably not sufficient to fix D with liability for the armed robbery and homicide which the son commits using one of D's guns. There was no intention to do any act of advice or assistance. 384 SANFORD H. KADISH [Vol. 87 ulated penalty? I ask that question on the level of principle for I would not think any radical extension of the criminal law prudent in today's world, at least in America. We already greatly overuse it in inappropriate areas and use it very poorly in appropriate areas. Moreover, the general concern that liability for reckless complicity would adversely impact ordinary affairs has some basis in reality even if I am correct that the concern is overstated. So prudence would suggest a preference for ad hoc statutes that punish reckless aiding of another's crime only in nar- rowly defined situations where the case for criminalization seems com- pelling; in child abuse cases, for example.37 But putting aside what is immediately prudent in today's world, I propose to raise some doubts about the seriousness of the concern that an enlarged liability for reckless complicity (even though restricted to crimes of recklessness) would unduly chill the conduct of ordinary affairs. My purpose is not to advance the cause of reckless complicity but to try to locate the deeper reasons for its general rejection in the law. We do, after all, accept it as right and natural to hold people criminally liable for harms they recklessly "cause." Why not also for harms they recklessly help or influence others to commit? Consider the Model Penal Code. It makes it criminal to recklessly kill another,38 to recklessly endanger another person,3 9 to recklessly risk a catastrophe, 4° to recklessly destroy property.4' But it does not make it criminal to recklessly occasion any of these harms through the actions of another. Why do we draw back, out of concem for creating an uncertainty in ordinary affairs, when the same harm is occasioned through another person's action? One possible reason may be this: since actions depend on human intentions, which are always subject to be made and remade at the will of the actor, there is not enough predictability in human affairs to make a rule of liability safe. But is that so? If the standard of liability were simple negligence or knowledge of a mere risk, that would be so (although it would also be so in the case of causing harms). But if the standard of recklessness were drawn with the kind of stringency exemplified by the influential Model Penal Code formulation it is not so clear. That standard, applied to a rule of liability for helping or en-

37 See supra note 35. Even in this situation, however, imposing liability on the mother may be excessive in light of the dependency and powerlessness of many defendants. See Dorothy E. Roberts, Motherhood and Crime, 79 IowA L. REv. 95 (1993). 38 MODEL PENAL CODE § 210.3 (1985). 39 Id. § 211.2. 40 Id. § 220.2. 41 Id. § 220.3. 1997] RECKLESS COMPLICITY couraging the criminal actions of others, would require that the actor "consciously disregard[ ] a substantial and unjustifiable risk" that the other person would commit the crime, the risk being "of such a nature and degree that, considering the nature and purpose of his conduct and the circumstances known to him, its disregard involves a gross deviation from the standard of conduct that a law-abiding person would observe in the actor's situation."42 It may be true that on the whole we can more safely predict natural happenings than human actions. Yet where a jury could conclude beyond a reasonable doubt that in a particular case the risk that defendant's action would help or encourage another's crime was substantial and unjustifiable and was known to be so by the defendant, one might reasonably think that liability could be safely imposed. There is a further consideration that tends to raise a doubt about the role of the policy argument in supporting the intention requirement. That requirement applies where the issue is the mens rea of S as to the adion of P; that is, S must act in order to encourage or help P do the action constituting the crime. But where the issue is the mens rea of S as to some circumstance of P's action, rather than the action itself, there is respectable support in England and the United States for settling for whatever mens rea suffices for the crime committed by P.43 An example: S lends P and P's underage girlfriend the keys to his apartment, which they use to have sex. S is liable for statutory rape if he gave them the keys in order for them to have sex, but not if he was just aware of a substantial risk that they would do so. This is the traditional position. Explanation? We don't want to make it too risky for people to do ordinarily lawful things because of what some other person might do. But suppose that though he did give them the keys in order that they might have sex he was unaware that the girl was under-

42 Id. § 2.02(2) (c). 43 For England, see Carterv. Richardson [1974] Road Traffic Reports 314 and Professor John C. Smith's discussion of it in Commentary, 1991 GRiM. L. REv. 765, as well as Draft Code Clause 27 which was intended to state existing law: A person is guilty of an offence as an accessory if - (a) he intentionally procures or assists or encourages the act which constitutes or results in the commission of the offense by the principal; and (b) he knows that, or (where recklessness suffices in the case of the principal) is reckless with respect to, any circumstance that is an element of the offence.... A Criminal Code for England and Wales, Law Comm. No. 177 (1989). Glanville Williams also supports the distinction, although he doubts that there is clear English authority on the subject. See Williams, supra note 27, at 98. For America the authority is less clear, although the Model Penal Code, while insisting on strict intention as to P's action, was prepared to accept for the circumstance element of P's crime whatever the crime required of P. Model Penal Code § 2.06 cmt. at 311 n.7 (1985). SANFORD H. KADISH [Vol. 87

age. He would be liable anyway, because no mens rea is required of him beyond that required of P. But why? Lending keys to a friend for a sexual encounter is ordinarily a lawful action. It is here unlawful only because the girl is underage. But S did not know that-perhaps he was reckless, or negligent, but no more. Isn't the upshot just the same as making recklessness enough for helping or encouraging P's action, namely to burden ordinary lawful actions with the risk of guessing right as to the crimes others may commit as a consequence of your action? I am not suggesting that S should not be liable. Statu- tory rape does not require knowledge of the age of the girl; recklessness, negligence, even strict liability suffices. So it makes sense to require no greater mens rea for S than is required for P. But that is just the argument I am making concerning the mens rea as to the action-if the crime of P requires no more than recklessness, then it makes the same sense to hold S liable for recklessly helping or encouraging him to do the action. Of course, there is a problem of vagueness with the standard of recklessness, since it calls for an evaluation of the substantiality and justifiability of the risk. And like all vague prohibitions it might tend to cast a shadow beyond its terms as people strove to reduce their exposure to criminal penalties. But we live with precisely that vagueness and the shadow it casts in the case of recklessly caused harms and in all cases where recklessness is the standard of liability as to some circumstance. As Justice Holmes declared, "the law is full of instances where a man's fate depends on his estimating rightly, that is, as the jury subsequently estimates it, some matter of degree."44 And in a later case the Supreme Court rejected a vagueness challenge to a crime defined in terms of reasonableness, stating: The mere fact that a penal statute is so framed as to require a jury upon occasion to determine a question of reasonableness is not sufficient 4to5 make it too vague to afford a practical guide to permissible conduct. If that is so with respect to recklessly caused harms, why should it not be so with respect to recklessly assisted harms as well? Consider some cases. Earlier I referred to Professor Williams' attack on a proposal for a general rule of reckless complicity. He of- fered the following hypotheticals to convey the unacceptability of such a rule: It would be unacceptable to hold that a man who lets a car on hire to another becomes a party to the latter's drink-driving, or to arson committed with the aid of the car (neither act being any part of his own purpose), if he is merely aware of the possibility (including probability,

44 Nash v. United States, 229 U.S. 373, 377 (1912). 45 United States v. Ragen, 314 U.S. 513, 523 (1941). 1997] RECKLESS COMPLICTTY

but not virtual certainty) of the car being used in this way by the borrower; or that a taxi-driver can become a criminal merely because he suspects that his passenger may be on a criminal expedition; or that a barman or social host is an accomplice in drink-driving if he is merely aware of the possibility of the customer or guest driving home with an excessive blood-alcohol count.. 46 True. But suppose the car owner is not merely aware of a possibility that the borrower will use the car to commit arson, but is conscious of a substantial and unjustifiable risk that the borrower would do so, a risk so great that its disregard involves a gross deviation from the standard of conduct a law-abiding person would observe. Or suppose the taxi-driver does not merely suspect the passenger may be on a criminal expedition, but is conscious of a substantial and unjustifiable risk that the passenger will do so. Or that the barman or social host is not merely aware of a possibility that the customer or guest will engage in drunk driving, but is conscious of a risk, again substantial and unjustifiable, that she will do so. It is not evident to me that subjecting actors in these circumstances to liability for a crime of recklessness need greatly imperil the security of otherwise lawful activities, certainly not any more than holding actors liable for recklessly "causing" harms, which the law regularly does. People aren't all that unpredictable. There is at least some recent authority that English law has already moved in this direction. In Blakely, Sutton v. Director of Public Prosecutions,47 the defendant, while sharing a drink with T at a pub, laced T's tonic water with vodka intending to disclose what she had done so that T would stay the night with her rather than drive home with alcohol in his blood. But T soon thereafter went to the toilet and unexpectedly left directly from there in his car so that she never had a chance to tell him. T was convicted of driving with excessive alcohol and defendant was convicted of procuring that offense. The court reversed because the trial court had assumed that an intention that T should commit the crime was not required for procuring a conviction. While the court held this to be error, it ventured the view that if the charge had been aiding and abetting it would have been sufficient that the defendant was reckless, in the sense of having knowingly taken a risk that T might drive with excessive alcohol in his blood. Professor Ashworth says that this case represents a "great and new- found width in the law of complicity."48 For reasons I have suggested,

46 Williams, supra note 27, at 101. 47 [1991] Road Traffic Reports 405. The case is reported and commented upon by Professor John C. Smith at 1991 CIuM. L. Rsv. 763, 765-67. 48 ANDRmv ASHWORTH, PRINCIPLES OF CRIMINAL LAW 424 (2d ed. 1995). 388 SANFORD H. KADISH [Vol. 87

I believe it is in principle a defensible extension. It must be acknowledged that a broad reckless complicity statute would criminalize some actions that are much less likely candidates for criminalization. Suppose, to vary the hypotheticals I posed two paragraghs earlier that in the first case a person, fully aware of the arsonist's intentions, shows him how to affix the car's seat belt; or in the second case that the person tells him where to wait for a taxi; or in the third case that a person simply cleans the glass for the inebriated guest. The problem in these cases, as the Model Penal Code pointed out when originally proposing to make knowing aid sufficient for complicity, is that the aid, which could so readily have been obtained elsewhere, makes too insignificant a contribution to the crime to justify the interference with otherwise lawful conduct.49 To address this problem, the Model Penal Code proposed that knowing aid be criminal only when it "substantially facilitates" P's crime.50 While this is a vague line, there is probably none less vague, as the Model Penal Code Comments observes, "that both affirms a liability without a purpose to facilitate the crime and gives the court and jury a discretion to avoid it when its imposition would be deemed extreme."51 A like qualification of any reckless complicity statute would be equally appropriate to help confine the scope of threatened criminality within acceptable limits. Still another qualification of reckless complicity which would serve to allay concerns over its overreach would confine its scope to reckless assistance of more serious crimes, felonies, for example, and continue to require purpose for misdemeanors or lesser offenses. That move has been proposed as a qualification to a doctrine of knowing complicity.52 It would serve the same purpose for a doctrine of reckless complicity. So, taken all in all, it seems to me that concern not to imperil the security of ordinary affairs is not a sufficient ground for rejecting a doctrine of reckless complicity even where the defendant's conduct is otherwise lawful. But before proceeding to the considerations that I believe lie at the root of the law's inhospitality to any doctrine of enlarged reckless complicity, I pause to consider the more technical question of how such a doctrine could be imported into the law if it were thought desirable to do so (which, as I said, I do not, for differ-

49 MODEL PENAL CODE § 2.04(3) cmt. at 30-31 (Tentative Draft No. 1 1953). 50 Id. at 31. 51 Id. But see K.J.M. Smith, The Law Commission ConsultationPaper on Complicity: (1) A Blueprintfor Rationalism, 1994 CRIM. L. REv. 239, 246-47. 52 See United States v. Fountain, 768 F.2d 790, 798 (7th Cir. 1985); People v. Lauria, 59 Cal. Rptr. 628, 635 (Cal. CL App. 1967). 1997] RECKLESS COMPLCTY

ent reasons). Would it accomplish its purposes if the doctrine were engrafted onto existing criminal codes? Yes, when P's crime can be committed recklessly: since in these cases S was reckless in helping P commit the crime, S himself can be held guilty of P's crime, or so I have argued. But suppose P's crime requires a greater culpability than recklessness? If S, lacking the required culpability for P's crime, cannot be convicted of it, then of what crime can he be convicted? Surely of a lesser version of P's crime that requires only recklessness-manslaughter rather than murder, for example. But generally there is no lesser crime of recklessness.53 Suppose in Hyde, for example, one of the other assailants, known by Hyde to be a notorious thief, took the occasion of the joint assault to lift the victim's wallet. If not robbery, there is no crime involving taking the wallet for which Hyde could be convicted. Or consider again the hypothetical of the person who revealed a safe's combination to a known career burglar in reckless disregard of the risk he would use the information to effect a burglary. A doctrine of enlarged reckless complicity would find no purchase since there would be no lesser crime of recklessness for which to convict her. This difficulty could be solved by a separate crime, with its own punishment, for one who recklessly helps or encourages another to commit a crime. Insofar as complicity is conceived as a way of convicting one person of the crime committed by another, such a statute would constitute an alternative to complicity rather than an expansion of it. But insofar as complicity is seen more broadly as a way of holding a person accountable for those consequences of his action that take the form of the crimes of another, a reckless complicity statute may fairly be regarded as a doctrine of complicity. There is a precedent for a solution of this kind in the criminal facilitation statutes enacted in a number of states that insist on intentional, rather than just knowing, aid or encouragement to the crime of another as the basis for accomplice liability.54 The effect of these statutes is to make the provision of knowing aid a separate crime of criminal facilitation with punishment set at some fraction of that provided for the crime committed. Insofar as these statutes reach only knowing aid they don't reach restricted reckless complicity, but the

53 It is sometimes possible to convict S of a lesser intentional crime than P's, but these are cases where S would be guilty of P's crime except for mitigating factors personal to him-an emotionally distraught and provoked S who hires a professional killer, for example. See Kadish, supra note 1, at 339-40. 54 See, e.g., ARIz. REv. STAT. ANN. § 13-1004 (West 1989); Ky. REv. STAT. ANN. § 506.080 (Michie 1990); N.D. CEN-r. CODE ANN. § 12.1-06-02 (Michie 1985). SANFORD H. KADISH [Vol. 87

New York version seems to go further by defining facilitation as providing a person with the means or opportunity to commit a crime, believing it probable that he is rendering aid to a person who intends to commit a crime.55 If "believing it probable" is taken as equivalent to recklessness, we have a statute establishing the enlarged reckless complicity doctrine in its restricted form. Even so interpreted, the New York statute is still short of captur- ing all the situations it might. It doesn't reach cases where the crime recklessly aided is itself one of recklessness (like our father and son hypothetical, or the drag race cases) and it doesn't reach cases where the means or opportunity are given to any who take them rather than to someone who the actor believes intends to commit a crime (like the case of a police officer who leaves his loaded gun in a schoolyard). A version of the New York statute enlarged to include these situations would come close to a reckless complicity doctrine. 56 As I said earlier, because of the pervasive overuse of the criminal law in America I do not favor such a statute. But that reason aside, what stands against it? Of those considerations that I suggested lay behind the requirement of intentionality, culpability is not a problem for a doctrine of reckless complicity, since it applies only to crimes for which S has the required culpability, namely recklessness. As for the policy concern that ordinary affairs not be jeopardized, I have earlier said why I believe that it does not justify the insistence on intention: if penalizing recklessly caused harms does not jeopardize ordinary affairs it is not clear why it is thought that penalizing recklessly assisted harms would. I suggest that the full explanation of the law's unrecep- tivity to reckless complicity derives from the third consideration I ad- vanced as supporting the requirement of intentionality, namely, an ethic of individualism and self-determinism.

V. THE ETHIC OF INDIVIDUALISM AND SELF-DETERMINISM In my references so far to individualism and self-determinism I have treated them as together constituting an ethic. But it should be observed that they contribute to that ethic in somewhat different ways. As I am using the terms, individualism is primarily a normatively

55 N.Y. PENAL LAw § 115.00 (McKinney 1987): A person is guilty of criminal facilitation . . . when, believing it probable that he is rendering aid.., to a person who intends to commit a crime, he engages in conduct which provides such person with means or opportunity for the commission thereof and which in fact aids such person to commit a felony .... 56 Not altogether, because the New York statute only reaches reckless aid, not reckless encouragement. But the concern for a possible chilling effect on lawful conduct is more serious in the case of words, a consideration that apparently prompted the restriction of the statute to cases of aid. 1997] RECKLESS COMPLICITY

driven perspective, self-determinism a conceptually driven one. I first deal with individualism. The reason for the law's insistence on intention is not simply that liability for recklessly assisting the crimes of others would be bad public policy in so far as it might threaten the security of lawful conduct. It is also, and I think more importantly, attributable to the individual- istic ethic, the belief that people's freedom to act within the law should not be restrained by considerations of wrongs others might commit. This, I suggest, more fully accounts for the law's different treatment of recklessly caused harms and recklessly assisted harms. A person may fairly be held accountable for the former; he caused them, they are what he did. The latter are what someone else did, not he, and he therefore should not be held responsible for them, unless he made them his own by intentionally helping another commit them. The treatment of criminal liability for omissions in English and American law offers an instructive parallel. The commonly stated ground for not punishing a person for failing to prevent harm to others that the person did not cause or have a duty to prevent is the insecurity such a rule would create in the everyday lives of ordinary citizens: how much of their own interests must people sacrifice in order to prevent how much harm to others that is not their doing? Since no line could be drawn with acceptable clarity, ordinary affairs would be rendered insecure and, in the words of Lord Macaulay, "the whole order of society" would be disturbed. 57 But it is evident that this can only be a partial explanation of the common law's limitations on liability for omissions. After all, the whole order of society is main- tained even though people are held liable for harms to others they recklessly cause by their actions. Why should it be otherwise for harms they could but do not prevent? Surely it is a problem to determine just how much of one's own interest need be sacrificed for how much of another's. But the same basic problem is confronted in determining what constitutes reckless action and is solved by the test of gross deviation from the standard of conduct of a reasonable law-abiding

57 Lord Macaulay and other Indian Law Commissioners, Indian Penal Code, Note M, in THE COMPLETE WORKS OF THOMAS BABINGTON MACAULAY: SPEECHES AND LEGAL STUDIES 651, 655 (University Ed. 1900): It is true that the man who, having abundance of wealth, suffers a fellow creature to die of hunger at his feet, is a bad man ..... But we are unable to see where, if we make such a man legally punishable, we can draw the line. If the rich man who refuses to save a beggar's life at the cost of a little copper is a murderer, is the poor man just one degree above beggary also to be a murderer if he omits to invite the beggar to partake his hard earned rice? Again, if the rich man is a murderer for refusing to save the beggar's life at the cost of a little copper, is he also to be a murderer if he refuses to save the beggar's life at the cost of a thousand rupees? SANFORD H. KADISH [Vol. 87

person.58 An analogous test for omissions would perhaps be some- thing like "gross deviation from the standard of common decency." Open-ended, but no more so than the standard of reasonableness. What is at work in omissions, then, is the norm of individualism which, while not necessarily denying the virtue of social responsibility, rejects the rightness, not just the imprudence, of coercing it by law. It is this same individualist ethic that strongly influences the traditional reluctance of the law to impose punishment on a person for harms that others cause by actions he does not intentionally assist, even though on a wider communal ethic he may be thought responsible for those actions. Now for the related self-determinist component of the ethic, which I suggested was primarily conceptual. I can best illustrate this by returning to the earlier discussion of the father who recklessly lends his car keys to his daredevil son, knowing he has not learned to drive, resulting in a fatal accident. As I said earlier most would feel that the father deserves less punishment than the son. Why should this be? It may be thought that the answer is the lesser probability of the harm happening as a result of his action than as a result of the son's action. For after all, whatever the probability of the son causing a fatal accident, the probability of the father's action resulting in a death is greater since it depends on the further contingency that the son should drive recklessly. But it is curious that we do not react in the same way to differing probabilities when the issue is causation, that is, when contingencies of happenings rather than of human actions in- tervene between the reckless action and the harm. Why not, if it is probabilities that are determining our sense of deserved punishment for unintended harms? An example: a driver knowingly drives with bald tires and bad brakes over a mountain road in winter, skids and kills someone. Do we at all feel his culpability is diminished because the accident depended on (i) a sudden rain storm, (ii) a drop in tem- perature which froze the rain on the road, and (iii) the presence of a hitchhiker on the shoulder of the road? I think not. Nor would the intervening events lead a court to deny liability because of the absence

58 Herbert Wechsler &Jerome Michael, A Rationaleof the Law of Homicide 37 COLUM. L. REv. 701, 751 n.175 (1937): Whereas the issue there is ... whether or not the act is a sufficiently necessary means to sufficiently desirable ends to compensate for the risk of death or injury which it creates, the issue here is whether or not freedom to remain inactive serves ends that are sufficiently desirable to compensate for the evil that inaction permits to befall. The extent of the burden imposed by the act is obviously a relevant factor in making such an evaluation. If the burden is negligible or very light, the case for liability is strong, and the difficulty of formulating a general rule no more insuperable an obsta- cle than in the case of acts. 1997] RECKLESS COMPLCIYT9 393 of proximate cause: a typical judicial response would be to find that the intervening events were "foreseeable." Yet the requirement of the concurrence of each of these events reduced the probability of the accident. I suggest that the differences in our reactions when it is a voli- tional human action that intervenes is not the product of perceived differences in probabilities, but of the pervasive conviction, widely manifested in the law, that it simply matters whether the causal route goes through another person, because we perceive human actions as differing from all other natural events in the world.5 9 This is what I mean by the conceptual pull of the self-determinist perspective. It is the same way of seeing the world that lies behind novus actus in- terveniens, human actions cause, they are not caused.60 So recklessness with respect to a natural happening is not seen to be commensurable with recklessness with respect to another person acting in a certain way. You may be as culpable as another for the harm the other causes if you exercise your will to participate in his action. Otherwise, what he causes is his doing, not yours. The reckless helper does not cause or participate in causing the harm done by the reckless doer. So even if we are prepared to find the helper blameworthy for his reckless contribution to the upshot we are inclined to see his culpability as necessarily less than that of the doer.6' As I say, I do not find all this convincing on a rational level, but it does seem to resonate with an ingrained view of the world which makes the causal route through which an upshot occurs a central feature in assessing blame. Where this view comes from is a bit mysteri- ous. I suspect it's a product of our evolutionary development, somewhat like the urge for retribution or the feeling that one who intends a harm deserves less punishment if he fails than if he succeeds.62 But that's speculation. What I feel more confident about is that the tension exemplified here between the moral distinctions we intuit and those we feel able rationally to defend is a common feature of the landscape of our moral experience and plays an important role in the shaping of our institutions, like the criminal law, which rests upon it. There remains, of course, the more controversial question of

59 Kadish, supra note 1. 60 Glanville Williams, Finis For Novus Actus?, 48 CAMBIDGE LJ. 391, 391-92 (1989). 61 This seems not to be so in common purpose or joint enterprise cases, but in them S does intend to assist P's crime, although a different one. This fact, combined with the pervasive retributive sentiment that one who embarks on a crime should be held for whatever ensues, see supra note 19, may explain the difference. 62 Sanford H. Kadish, Foreword: The CriminalLaw and the Luck of the Draw, 84J. CRIM. L & CRIMINOLOGY 679, 679 (1994). SANFORD H. KADISH [Vol. 87

whether we should strive to conform law to the instinctive or to sup- press it as far as we can in favor of the rational. Justice Holmes once said that "[t]he law can ask no better justification than the deepest instincts of man."63 I would hope it could ask for more. Certainly it is generally desirable that laws be so grounded, and it may be to some degree inevitable. But "no better justification," if more than a Holme- sian rhetorical flourish, reflects either an uncommonly charitable view of human instincts, or, more likely in Holmes' case, a gloomy ethic that identifies the normative with passion and instinct and downplays the role of reason and critical morality.64 In contrast to Holmes' Dio- nysian vision John Stuart Mill offers an Apollonian one: [N] early every respectable attribute of humanity is the result, not of instinct, but of a victory over instinct.... [T]he duty of man is the same in respect to his own nature as in respect to the nature of all other things, namely, not to follow but to amend it.... [While we have good instincts] it must be allowed that we have also bad instincts which it should be the aim of education not simply to regulate but to extirpate, or rather ...to starve them by disuse.65 Mill strikes the more appealing and stirring chord, but the Holmes perspective at least provides the note of caution that law has practical work to do and must govern people as they are, and that it can not stray too far from deep rooted common perceptions without under- mining public acceptance and compliance.

63 Oliver Wendell Holmes, The Path of the Law, in COLLECTED LEGAL PAPERS 167, 200 (1920). 64 For a defense of conforming principles of criminal liability to popular intuitions revealed by community survey research methods, see PAUL H. ROBINSON AND JOHN M. DARLEY, JUSTICE, LIABILry, AND BLAME: COMMUNITY ViEWS AND THE CuMINAL LAW (1995). 65 John Stuart Mill, Nature in EssENrIAL Woxs OFJOHN STuART MILL 367, 390, 395, 396 (Max Lerner ed., 1965).